Select the column with at least one 1 in it

Hey people

Assume I have a m by n matrix of binary data. How would I be able to identify the column that has the least number of ones?

I used the sum function and compared the max and min. Later on in many loops I realised that it also identifies the column with zero 1s too. How do I avoid that?

The column to be selected must have a minimum of one 1s to n number of ones.

b = sum(H);
bmax = max(b);
bmin = min(b);

RC = find(b==min(b));
%This is what I am using which identifies the column with zero ones. Any simple methods?

Answer 1

[Walter Roberson]

find(any(H,1))

If you need at most n ones then:

b = sum(H,1);
find(b >= 1 & b <= n)

[Rishi Balasubramanian]

okay. Now if i am only looking at a specific row column range.

If the matrix is 20x30 and I need to look for the column with the least number of 1s from row 1:15 and column 1:25, how do I do that?

I tried your code find(any(H,1)) and it reads 1s across the matrix.

[Walter Roberson]

any(H(1:15,1:25),1)

[Rishi Balasubramanian]

Consider a matrix H = 1x30.

The data in those elements are numbers 0, 1, 2, ......, n

I wanna be able select the column(s) with the values from 1.

If H has a '0', then columns having '1' must be returned.

If H has a '0', and has no 1s, then columns having '2' must be returned.

If H has a '0', and has no 1s and 2s, then columns having '3' must be returned.... and so on.

It should be able to return all the columns with a minimum value in it, where the minimum value cannot be zero. But it is any available minimum value after 0.

How do i achieve this?

[Walter Roberson]

any(H==min(H(H~=0)),1)